- From: Toby Inkster <tai@g5n.co.uk>
- Date: Sun, 15 Nov 2009 18:44:23 +0000
- To: Shane McCarron <shane@aptest.com>
- Cc: Ivan Herman <ivan@w3.org>, "public-rdf-in-xhtml-tf.w3.org" <public-rdf-in-xhtml-tf@w3.org>
On Sat, 2009-11-14 at 16:28 -0600, Shane McCarron wrote: > I agree that the technique described will work. I am surprised that > you think it ok okay to change the bnode in the emitted triples. While > you might not think it has a meaning, others may... The RDF Semantics[1] recommendation says this about identifiers for blank nodes: > While node identifiers such as '_:xxx' serve to identify blank nodes > in the surface syntax, these expressions are not considered to be the > label of the graph node they identify; they are not names, and do not > occur in the actual graph. i.e. node identifiers are syntactic sugar. They're a very useful bit of syntactic sugar in fact, because without them it would be pretty much impossible to serialise a graph containing cyclical references between blank nodes. But they're not a part of the RDF graph itself - so, assuming that all an RDFa processor is required to output is an RDF graph, it cannot be a requirement for RDF processors to retain blank node identifiers. That said, retaining blank node identifiers when possible is a good thing - it helps the readability of any resultant serialisation, in much the same way as retaining CURIE prefixes of the input RDFa does. The method I outlined does retain blank node identifiers in the vast majority of cases, and can be guaranteed not to produce collisions, yet it's still possible to implement as part of a one-pass algorithm. ____ 1. http://www.w3.org/TR/2004/REC-rdf-mt-20040210/ -- Toby A Inkster <mailto:mail@tobyinkster.co.uk> <http://tobyinkster.co.uk>
Received on Sunday, 15 November 2009 18:45:14 UTC